Analysis of Correlation Between Whole Exome Sequencing and Ultrasound Examinationin Prenatal Diagnosis of Fetal Skeletal Dysplasia

Background:Fetal skeletal dysplasia is a disease that is dicult to distinguish these types of diseases during the fetal period. Due to the diculty of fetal ultrasound diagnosis, the severity of fetal skeletal dysplasia is extremely dicult to assess. Methods: 79 fetal samples of skeletal dysplasia from the third aliated hospital of Zhengzhou University, China from August 2018 to April 2020, which had undergone prenatal whole exome sequencing(WES) were analyzed and the results of whole-exome sequencing and fetal ultrasound test results were compared. Results:We nd that the fetal short limb phenotype found in the range of FL<-4.0SD or HL<-4.0SD through ultrasound test is closely related to FGFR3 gene mutation , and the correlation is stronger when accompanied by macrocephaly. We also nd that the fetal limb curved phenotype is closely related to COL1A1 gene mutation. At the same time, we nd that nasal dysplasia during fetal period is also a common phenotype of abnormal results detected by whole exome sequencing. Conclusions: Our research shows that WES has different detection rates for various skeletal abnormalities according to the different types of ultrasound detection results, which provides a meaningful guidance for clinical diagnosis of fetal skeletal dysplasia.


Introduction
Fetal skeletal dysplasia is a disease of osteochondrocytes with strong clinical variability [1,2]. Due to their rarity, it is di cult to distinguish these types of diseases during the fetal period [3]. The diagnosis of fetal skeletal dysplasia can be done through prenatal ultrasound evaluation [3]. Skeletal dysplasia often causes a reduction in bust size and leads to lung dysplasia, so fetal skeletal dysplasia is often fatal. However, due to the di culty of fetal ultrasound diagnosis, the severity of fetal skeletal dysplasia is extremely di cult to assess [4]. The accuracy of conventional ultrasound on fetal skeletal dysplasia is no more than 40% [4]. Misdiagnosis can lead to erroneous information about the risk of relapse and delay optimal fetal management.
Using fetal imaging as a guide, a personalized prenatal genetic examination strategy can be selected. Current options include chromosome karyotype analysis, chromosome uorescence in situ hybridization experiments, chromosome microarray analysis, whole exome sequencing etc [5]. According to the 2010 revision of the Nosology and Classi cation of Genetic Skeletal Disorders, 456 diseases were divided into 40 groups according to molecular, biochemical and radiological standards [6]. Among them, 316 bone-related diseases are related to the mutation of one or more genes in 226 different genes, which provides a basis for the molecular genetic diagnosis of fetal skeletal development abnormalities [6]. Two recent studies have shown that in the assessment of large scale prospectively ascertained cohort of fetal with skeletal anomalies, the diagnostic rates of skeletal dysplasia are 10/65 (15.4%) [7] and 8/34 (24%) [5], respectively.
In our study, we analyzed 79 fetal samples of skeletal dysplasia from the third a liated hospital of Zhengzhou University, China from August 2018 to April 2020, which had undergone prenatal whole exome sequencing(WES). Our research aims to explore the genetic types of fetal skeletal dysplasia in China, and explore the correlation between results by WES and phenotype of skeletal dysplasia in the fetal period by ultrasound, with a view to providing a theoretical basis for early implementation of birth defect intervention and reproductive risk assessment.

Results
Correlation between positive results by WES and clinical phenotypes by ultrasound Among 79 pregnant women from the third a liated hospital of Zhengzhou University, China, from August 2018 to April 2020, 25 cases that were positive by WES were detected,whose detection rate was 31.6%.Among these cases, the detection rate of cases with only FL<-4.0SD or only HL<-4.0SD was 41.6% (5/12).
The detection rate of cases with both FL<-4.0SD and HL<-4.0SD is 60% (3/5).In cases of -4.0SD<FL<-2.0SD or -4.0SD<HL<-2.0SD, the detection rate by WES was 29.4% (5/17).However,in these cases where exsists -4.0SD<FL<-2.0SD or -4.0SD<HL<-2.0SD, with long bone bending, the detection rate by WES was 100% (2/2). In cases of FL<-2.0SD or HL<-2.0SD, the detection rate by WES is 20% (1/5), where the only one positive case besides FL<-2.0SD and HL<-2.0SD, there exsists also femoral curvature. In cases with only long bone bending, the detection rate by WES is 50% (2/4).In cases of nasal bone dysplasia, the detection rate by WES is 80% (4/5).It is worth mentioning that cases of only microcephaly have not been detected positive by WES. Only 1 of 9 cases of fetal hand or foot deformity were positive by WES (11.1%). (Table .1) Correlation between abnormal results of FGFR3 gene sequenced by WES and clinical phenotypes by ultrasound Among25 cases that are positive by WES, 7 cases are caused by FGFR3 gene mutation, accounting for 28%.Among them, 6 cases are FGFR3 gene c.1138G>A mutation, and 1 case is FGFR3 gene c.1620C>A mutation.In these 7 cases, the ultrasound test results of almost all cases show FL<-4.0SD or HL<-4.0SD(Only one case of short limbs with ambiguous data),among which 3 cases of ultrasound test results also show macrocephaly, accounting for 42.8%.Besides, all genetic variations are de novo. (Table.2) Correlation between abnormal results of COL1A1 gene sequenced by WES and clinical phenotypes by ultrasound Among25 cases that are positive by WES, 3 cases are caused by COL1A1 gene mutation, accounting for 12%. In these 3 cases, the ultrasound test results of 2 cases show FL>-4.0SD or HL>-4.0SD, ),among which all of 3 cases of ultrasound test results showfetal limbs slightly curved, accounting for 100%. Besides, One of the three cases has a COL1A1 gene variant inherited from the pregnant husband, and the rest of COL1A1 gene variant are de novo. (Table.3) Correlation between abnormal results of other gene sequenced by WES and clinical phenotypes by ultrasound Among25 cases that are positive by WES,11 cases are caused by other gene mutation accounting for 44%. In these 11 cases,there are 2 cases caused by RUNX2 gene mutation, both of which have nasal bones not shown by ultrasound.In addition, there are 2 cases due to COL2A1 gene mutation, both of which have short limbs by ultrasound.The other 7 cases are caused by ARSE gene mutation, EFTUD2 gene mutation, SCN4A gene mutation, COL1A2 gene mutation, PEX7 gene mutation, NEB gene mutation and FGFR2 gene mutation. Ultrasound test results related to these 7 genetic variations are mainly manifested by phenotypes other than short limbs, such as microcephaly, bilateral foot varus, the femur is slightly curved, abnormal ossi cation of both humerus and femur, ngers merge , toes merge, etc. (Table.

Discussion
By studying the WES test results of 79 pregnant women from the third a liated hospital of Zhengzhou University from August 2018 to April 2020, the overall detection rate of skeletal abnormalities was 31.6%. This is higher than the previously reported detection rate (15.4% [7] and 24% [5]). This may be due to our stricter requirements for the ultrasound test results of the enrolled cases. Interestingly, we nd that for different types of skeletal abnormalities, the detection rate through WES varies greatly. Among them, WES has a higher detection rate for short limbs. Especially when FL<-4.0SD or HL<-4.0SD, the detection rate can rise to 41.6%. However, when the fetal has short limbs with other bone abnormal phenotypes, the detection rate will be higher. For example, when the limbs are short with bone curved, the WES detection rate can reach 100%; when the limbs are short with nasal bone dysplasia, WES detection rate can reach 80%. On the contrary, if the phenotype of short limbs is not detected by ultrasound testing and only other skeletal abnormal phenotypes exists, such as only fetal hand or foot deformities, the WES detection rate will be very low, only 11.1% . It may be that achondroplasia is the cause for the most of skeletal abnormalities [8], which is also the most common form of inherited disproportionate short stature [9]. In our study, through ultrasound,46 of 79 fetal skeletal abnormalities had clinical manifestations of short limbs, accounting for 58.2%, consistent with previous reports [10]. This shows that the diagnosis of achondroplasia is the key to the diagnosis of fetal skeletal abnormalities.
Through our research, we nd that for the diagnosis of fetal FGFR3-related achondroplasia, the WES detection rate is highly correlated with the results of ultrasound testing of the fetal limb shortness and severity.Through our research, we nd that the diagnosis of fetal FGFR3-related achondroplasia [11] accounts for the highest proportion of all skeletal abnormalities, reaching 28%. Almost all the ultrasound test results of fetal achondroplasia have been detected with FL<-4.0SD or HL<-4.0SD, suggesting that FL<-4.0SD or HL<-4.0SD is the most critical basis for the diagnosis of achondroplasia by fetal ultrasound testing.At the same time, the detection of macrocephaly through ultrasound is also an important evidence for the diagnosis of fetal achondroplasia.Among the 7 cases of FGFR3-related achondroplasia, 6 cases were caused by FGFR3 gene c.1138G>A mutation. FGFR3 gene c.1138G>A mutation is the most common mutation in FGFR3-related achondroplasia, which accounts for more than 99% of all FGFR3-related achondroplasia together with FGFR3 gene c.1138G>C mutation [12,13].At this point, our research is consistent with these reports.In addition, we have also detected a case of FGFR3 gene c.1620C>A mutation through WES, which is related to hypochondroplasia [14].FGFR3 gene c.1620C>A mutation has been reported extensively [15] and is considered to account for 50%-76% of FGFR3-related hypochondroplasia [16].Therefore, FGFR3 gene c.1138G>A mutation and c.1620C>A mutation may be the two most common causes of skeletal abnormalities in the fetal period and require special clinical attention.
Osteogenesis imperfecta caused by the COL1A1 gene is also a major cause of fetal skeletal dysplasia. The main feature of Osteogenesis imperfecta is multiple fractures usually caused by minor trauma [17,18]. This may be manifested by the limbs being curved by ultrasound testing during the prenatal period.In our study, osteogenesis imperfecta due to COL1A1 gene mutation was detected in the presence of limbs slightly curved through ultrasound in the fetal period,which suggests that the presence of this phenotype during the fetal period may be highly correlated with COL1A1-related osteogenesis imperfecta.
In addition, COL1A1 gene c.896G>A mutation and c.1301G>A mutation found in our research have not been reported in other studies before, which broadens the clinical understanding of this gene mutation. COL1A1 gene c.3235G>A mutation, according to previous reports, shows that it has a highly variable phenotype in the family, and family members with this gene mutation can only show signs of disease without fractures [19,20]. This may explain why in our study, the husband of the pregnant woman also has the heterozygous mutation of the gene but the phenotype was not abnormal.
In our study, 2 cases of cleidocranial dysplasia related to RUNX2 gene mutation were detected, which were c.931_946del mutation and c.568C>T mutation.
The main clinical features of cleidocranial dysplasia include persistently open skull sutures with bulging calvaria, hypoplasia or aplasia of the clavicles permitting abnormal facility in apposing the shoulders, wide pubic symphysis, short middle phalanx of the fth ngers, dental anomalies, and often vertebral malformation [21]. It has been pointed out in previous reports that cleidocranial dysplasia can also be related to the phenotype of nasal bone loss [21], which is consistent with our case study.In our study, some other genetic variations or chromosomal aneuploidies related to the nasal bone loss phenotype are also found, such as the ARSE gene c.331C>T variation and trisomy 21.ARSE gene mutation is related to X-linked recessive chondrodysplasia punctata, which is manifested as nasal dysplasia and distal phalanx dysplasia [22], which is consistent with the phenotype observed during fetal ultrasound testing.Trisomy 21 is the most frequent form of mental retardation caused by a microscopically demonstrable chromosomal aberration, is characterized by well-de ned and distinctive phenotypic features and natural history [23]. It has been reported that nasal bone dysplasia is a common detected phenotype of fetal trisomy 21 [24], which is consistent with the phenotype of a trisomy 21 case found in our study. Therefore, in the diagnosis of fetal skeletal abnormalities, nasal bone dysplasia may be another typical indication of ultrasound abnormality in addition to short limbs.
In our study, some gene mutations related to skeletal abnormalities related to hand and foot abnormalities are also detected. For example, SCN4A gene c.4361G>A mutation [25], NEB gene c.1569+5G>A and c.2278C>T compound heterozygous mutation [26], FGFR2 gene c.755C>G mutation [27]. Most of the diseases related to these gene mutations mainly affect the function of the muscular system as the main cause [28,29], and their detection rate in fetal skeletal abnormalities is low (only FGFR2 gene c.755C>G mutation), but their detection can play an important guiding role in clinical diagnosis and treatment.

Conclusions
In conclusions, our research shows that the application of whole exome sequencing technology can signi cantly improve the systemic prenatal diagnosis of skeletal abnormalities, and according to the different types of ultrasound detection results, WES has different detection rates for various skeletal abnormalities. Through our research, it is shown that fetal short limbs are the best detection targets for WES to detect skeletal abnormalities. In addition, the fetal limbs curved and nasal bone dysplasia are also important clinical phenotypes that suggest genetic variation-related skeletal abnormalities. However, the genetic basis of bone diseases is still unknown in other respects, indicating that new genes or non-genetic factors may cause these diseases.

Materials And Methods
Pregnant women Enrollment specimens and DNA preparation We selected a total of 79 pregnant women from the third a liated hospital of Zhengzhou University, China, from August 2018 to April 2020, who had undergone whole exome sequencing due to fetal ultrasound suggesting skeletal dysplasia. Sample entry requirements: pregnant women between 18-36 weeks of gestation, ultrasound examination shows femur length <-2SD or humerus length <-2SD or other skeletal development abnormalities: such as nasal bone dysplasia, limb bone deformity, skull or clavicle development Exception etc.We performed amniocentesis on the enrolled pregnant women for subsequent DNA extraction.Afterwards, we used QIAGEN 69504 blood and tissue DNA extraction kit for DNA extraction from amniotic uid, peripheral blood of pregnant women and their husbands.

Whole exome sequencing(WES) and CNV analysis
The genomic DNA of amniotic uid was broken into random fragments and puri ed, and whole genome exome capture was performed for sequencing library preparation. The Illumina Hiseq XTen sequencer was used to perform double-ended high-throughput sequencing with a length of 150 bp. The raw data obtained by the sequencing were quality-controlled, and the basic data were analyzed and ltered to remove the linker sequence and repeat sequence. Data were analyzed using the Anno variant site detection system, the XYGeneRanger variant site annotation interpretation system.

Variant annotation and interpretation
The pathogenicity assessment and data interpretation rules of mutation are based on the guidelines of the American College of Medical Genetics and Genomics (ACMG) [30]. By querying 1000genomes, ExAC and gnomAD, we excluded gene mutations whose frequency of the population are more than 1% and removed non-functional mutation sites (such as synonymous mutations, non-coding region mutations, etc.).We performed gene function prediction (using software such as SIFT, Polyphen2, CADD, etc.) and clinical symptoms comparison. At last, we searched related disease databases and related references, and nally found candidate gene mutation sites for family veri cation. The variant annotation databases which were used include Human Genome hg19/GRCh37, RefSeq, dbSNP, 1000 Genomes phase3, ExAC, and gnomAD and the interpretation databases which were used include DGV, DECIPHER, OMIM, UCSC, ClinVar, HGMD and PubMed.
Family veri cation by sanger sequencing Genetic modi cation of the mutations was performed on the fetus, pregnant women and their husbands. According to the exome of the mutation site, the primers needed for sequencing the synthetic DNA fragments were designed, and the DNA of the fetus, pregnant women and their husbands were PCRampli ed. Method of sequencing was performed by Sanger sequencing and the sequencing results were compared with the results of whole exome sequencing. Declarations